Published online: January 15, 2018

Resource

Mapping protein interactions of sodium channel

NaV1.7 using epitope-tagged gene-targeted mice Alexandros H Kanellopoulos1,†, Jennifer Koenig1,†, Honglei Huang2, Martina Pyrski3 , Queensta Millet1, Stéphane Lolignier1,4, Toru Morohashi1, Samuel J Gossage1, Maude Jay1, John E Linley1,5, Georgios Baskozos6 , Benedikt M Kessler2, James J Cox1, Annette C Dolphin7, Frank Zufall3, John N Wood1,* & Jing Zhao1,**

Abstract Introduction

The voltage-gated sodium channel NaV1.7 plays a critical role in Pain is a major clinical problem. A 2012 National Health Interview

pain pathways. We generated an epitope-tagged NaV1.7 mouse Survey (NHIS) in the United States revealed that 25.3 million adults that showed normal pain behaviours to identify channel-inter- (11.2%) suffered from daily (chronic) pain and 23.4 million

acting proteins. Analysis of NaV1.7 complexes affinity-purified (10.3%) reported severe pain within a previous 3-month period under native conditions by mass spectrometry revealed 267 (Nahin, 2015). Many types of chronic pain are difficult to treat as proteins associated with Nav1.7 in vivo. The sodium channel b3 most available drugs have limited efficacy and can cause side (Scn3b), rather than the b1 subunit, complexes with Nav1.7, and effects. There is thus a huge unmet need for novel analgesics we demonstrate an interaction between collapsing-response medi- (Woodcock, 2009). Recent human and animal genetic studies have

ator protein (Crmp2) and Nav1.7, through which the analgesic drug indicated that the voltage-gated sodium channel (VGSC) NaV1.7

lacosamide regulates Nav1.7 current density. Novel NaV1.7 protein plays a crucial role in pain signalling (Nassar et al, 2004; Cox et al,

interactors including membrane-trafficking protein synaptotag- 2006; Dib-Hajj et al, 2013), highlighting NaV1.7 as a promising drug min-2 (Syt2), L-type amino acid transporter 1 (Lat1) and trans- target for development of novel analgesics (Habib et al, 2015; membrane P24-trafficking protein 10 (Tmed10) together with Emery et al, 2016). Scn3b and Crmp2 were validated by co-immunoprecipitation (Co- VGSCs consist of a large pore-forming a-subunit (~260 kDa) IP) from sensory neuron extract. Nav1.7, known to regulate opioid together with associated auxiliary b-subunits (33–36 kDa) and play receptor efficacy, interacts with the G protein-regulated inducer of a fundamental role in the initiation and propagation of action poten- neurite outgrowth (Gprin1), an opioid receptor-binding protein, tials in electrically excitable cells. Nine isoforms of a-subunits

demonstrating a physical and functional link between Nav1.7 and (NaV1.1–1.9) that display distinct expression patterns, and variable opioid signalling. Further information on physiological interactions channel properties have been identified in mammals (Frank &

provided with this normal epitope-tagged mouse should provide Catterall, 2003). NaV1.7, encoded by the gene SCN9A in humans, is useful insights into the many functions now associated with the selectively expressed peripherally in dorsal root ganglion (DRG),

NaV1.7 channel. trigeminal ganglia and sympathetic neurons (Toledo-Aral et al, 1997; Black et al, 2012), as well as in the central nervous system

Keywords NaV1.7; pain; protein–protein interactor; sensory neuron; sodium (Weiss et al, 2011; Branco et al, 2016). As a large membrane ion

channel channel, NaV1.7 produces a fast-activating, fast-inactivating and Subject Categories Molecular Biology of Disease; Neuroscience slowly repriming current (Klugbauer et al, 1995), acting as a thres- DOI 10.15252/embj.201796692 | Received 5 April 2017 | Revised 30 November hold channel to contribute to the generation and propagation of 2017 | Accepted 5 December 2017 | Published online 15 January 2018 action potentials by amplifying small sub-threshold depolarisations The EMBO Journal (2018) 37: 427–445 (Rush et al, 2007). The particular electrophysiological characteris-

tics of NaV1.7 suggest that it plays a key role in initiating action

1 Molecular Nociception Group, WIBR, University College London, London, UK 2 TDI Mass Spectrometry Laboratory, Target Discovery Institute, University of Oxford, Oxford, UK 3 Center for Integrative Physiology and Molecular Medicine, Saarland University, Homburg, Germany 4 Université Clermont Auvergne, Inserm U1107 Neuro-Dol, Pharmacologie Fondamentale et Clinique de la Douleur, Clermont-Ferrand, France 5 Neuroscience, IMED Biotech Unit, AstraZeneca, Cambridge, UK 6 Division of Bioscience, University College London, London, UK 7 Department of Neuroscience, Physiology and Pharmacology, University College London, London, UK *Corresponding author. Tel: +44 207 6796 954; E-mail: [email protected] **Corresponding author. Tel: +44 207 6790 959; E-mail: [email protected] † These authors contributed equally to this work

ª 2018 The Authors. Published under the terms of the CC BY 4.0 license The EMBO Journal Vol 37 |No3 | 2018 427 Published online: January 15, 2018

The EMBO Journal NaV1.7-interacting proteins Alexandros H Kanellopoulos et al

potentials in response to depolarisation of sensory neurons by (Fig 1B). The channel function of TAP-tagged NaV1.7 was examined noxious stimuli (Habib et al, 2015). In animal studies, our previous with electrophysiological analysis, and the result showed that all the

results demonstrated that conditional NaV1.7 knockout mice have cells presented normal functional NaV1.7 currents (Fig 1C). Activa-

major deficits in acute, inflammatory and neuropathic pain (Nassar tion and fast inactivation data were identical for wild-type NaV1.7

et al, 2004; Minett et al, 2012). Human genetic studies show that and TAP-tagged NaV1.7 (Fig 1D) demonstrating that the TAP-tag

loss-of-function mutations in NaV1.7 lead to congenital insensitivity does not affect the channel function of NaV1.7. to pain, whereas gain-of-function mutations cause a range of painful

inherited disorders (Cox et al, 2006; Dib-Hajj et al, 2013). Recent Generation of a TAP-tagged NaV1.7 mouse

studies also show that NaV1.7 is involved in neurotransmitter release in both the olfactory bulb and spinal cord (Weiss et al, 2011; We used a conventional gene targeting approach to generate an

Minett et al, 2012). Patients with recessive NaV1.7 mutations are epitope-tagged NaV1.7 mouse. The gene targeting vector was normal (apart from being pain-free and anosmic), suggesting that constructed using an Escherichia coli recombineering-based method

selective NaV1.7-blocking drugs are therefore likely to be effective (Bence et al, 2005; Fig EV1). A TAP-tag, which contains a HAT analgesics with limited side effects (Cox et al, 2006). domain, a TEV cleavage site and 3x FLAG-tags (Fig 1A), was Over the past decade, an enormous effort has been made to inserted into the open reading frame at the 30-end prior to the stop

develop selective NaV1.7 blockers. Efficient and selective NaV1.7 codon in exon 27 of NaV1.7 (NCBI Reference: NM_001290675; 0 antagonists have been developed; however, NaV1.7 antagonists Fig 2A). The final targeting vector construct containing a 5 homol- require co-administration of opioids to be fully analgesic (Minett et al, ogy arm (3.4 kb), a TAP-tag, neomycin cassette and a 30 homology

2015). NaV1.7 thus remains an important analgesic drug target and an arm (5.8 kb; Fig 2B) was transfected into the 129/Sv embryonic

alternative strategy to target NaV1.7 could be to interfere with either stem (ES) cells. Twelve colonies with the expected integration (tar- intracellular trafficking of the channel to the plasma membrane, or the geting efficiency was 3.5%) were detected by Southern blot analyses downstream effects of the channel on opioid peptide expression. (Fig 2C). Germline transmission and intact TAP-tag insertion after Although some molecules, such as Crmp2, ubiquitin ligase Nedd4-2, removal of the neomycin (neo) cassette were confirmed with South- fibroblast growth factor 13 (FGF13) and PDZ domain-containing ern blot, PCR and RT–PCR, respectively (Fig 2D–F). This mouse line TAP protein Pdzd2, have been reported to associate with the regulation of is henceforth referred to as NaV1.7 .

trafficking and degradation of NaV1.7, the entire protein–protein inter- TAP action network of NaV1.7 still needs to be defined (Shao et al,2009; NaV1.7 mice have normal pain behaviour Dustrude et al, 2013; Laedermann et al, 2013; Bao, 2015; Yang et al, TAP 2017). Affinity purification (AP) and targeted tandem affinity purifica- The homozygous NaV1.7 knock-in mice (KI) were healthy, tion (TAP) combined with mass spectrometry (MS) is a useful method fertile and apparently normal. Motor function of the mice was for mapping the organisation of multiprotein complexes (Angrand examined with the Rotarod test. The average time that KI animals et al, 2006; Wildburger et al, 2015). Using the AP-MS or TAP-MS stayed on the rod was similar to the WT mice (Fig 3A), suggest- method to characterise protein complexes from transgenic mice allows ing there is no deficit on motor function in the KI mice. The the identification of complexes in their native physiological environ- animal responses to low-threshold mechanical, acute noxious ment in contact with proteins that might only be specifically expressed thermal, noxious mechanical stimuli and acute peripheral in certain tissues (Fernandez et al, 2009). The principal aim of this inflammation were examined with von Frey filaments, Harg- study was to identify new protein interaction partners and networks reaves’ test, Randall–Selitto apparatus and formalin test, respec- involved in the trafficking and regulation of the sodium channel tively. The results showed that TAP-tagged KI mice had identical

NaV1.7 using mass spectrometry and our recently generated epitope responses to these stimuli compared to the littermate WT control TAP (TAP)-tagged NaV1.7 knock-in mouse. Such information should throw mice (Fig 3B–E), indicating NaV1.7 mice have normal acute

new light on the mechanisms through which NaV1.7 regulates a vari- pain behaviour. ety of physiological systems, as well as propagating action potentials,

especially in pain signalling pathways. Expression pattern of TAP-tagged NaV1.7 in the nervous system

We used immunohistochemistry to examine the expression pattern TAP Results of TAP-tagged NaV1.7 in the nervous system of NaV1.7 mice. The results showed that the FLAG-tag was expressed in the olfactory

The TAP-tag does not affect NaV1.7 channel function bulb, with strong staining visible in the olfactory nerve layer, the glomerular layer, and in the accessory olfactory bulbs (Fig 4A–D),

Prior to generating the TAP-tagged NaV1.7 knock-in mouse, we consistent with previous results (Weiss et al, 2011). In the brain,

tested the channel function of TAP-tagged NaV1.7 in HEK293 cells FLAG-tag expression was present in the medial habenula, the by establishing a HEK293 cell line stably expressing TAP-tagged anterodorsal thalamic nucleus, the laterodorsal thalamic nucleus

NaV1.7. A 5-kDa TAP-tag consisting of a poly-histidine affinity tag and in the subfornical organ that is located in the roof of the dorsal (HAT) and a 3x FLAG-tag in tandem (Terpe, 2003), separated by a third ventricle (Fig 4E and F) and is involved in the control of thirst unique TEV protease cleavage site, was fused to the C-terminus of (Oka et al, 2015). The FLAG-tag was also present in neurons of the

NaV1.7 (Fig 1A). The expression of TAP-tagged NaV1.7 was detected posterodorsal aspect of the medial amygdala and in the hypothala- with immunocytochemistry using an anti-FLAG antibody. The result mus in neurons of the arcuate nucleus (Fig 4G–I) as confirmed in a

showed that all the HEK293 cells expressed TAP-tagged NaV1.7 recent study (Branco et al, 2016). A clear staining appeared in

428 The EMBO Journal Vol 37 |No3 | 2018 ª 2018 The Authors Published online: January 15, 2018

Alexandros H Kanellopoulos et al NaV1.7-interacting proteins The EMBO Journal

A HAT Tag TEV site 3X FLAG Tag SRKDHLIHNVHKEEHAHAHNKIENLYFQGELPTAADYKDHDGDYKDHDIDYKDDDDK spacer

TAP tag

PCMV TAP-tagged SCN9A (6149 bp) pA

B An-FLAG DAPI Merged TAP-Nav1.7 HEK293

An-FLAG DAPI Merged HEK293

C HEK293 Ctrl

D Activation Inactivation NaV1.7 0 1.0 TAP-tagged NaV1.7 Na 1.7 25 ms -10 v 40 mV 0.8 -20

TAP-tagged NaV1.7 x 0.6 -30 Na 1.7 -120 mV v -100 mV -40 ma I 0.4

40 mV / 500ms 25ms I

I(pA/pF) -10mV -10mV TAP-tagged NaV1.7 -50 0.2 -60 -80 mV 0.0 -120mV -70 -130mV

200 pA -120 -100 -80 -60 -40 -20 0 20 40 60 -140 -120 -100 -80 -60 -40 -20 0 2 ms Vm(mV) Vm(mV)

Figure 1. Characterisation of TAP-tagged NaV1.7 in HEK293 cells.

A The diagram of TAP-tagged SCN9A cDNA construct used to establish the stably expressing TAP-tagged NaV1.7 HEK293 cell line. A sequence encoding a TAP-tag was

cloned immediately upstream of the stop codon of SCN9A coding the wild-type NaV1.7. B Left panel: Representative immunohistochemistry with an anti-FLAG antibody (in green) on HEK293 cells stably expressing TAP-tagged NaV1.7 (top panel) and parental HEK293 cells (bottom panel). Middle panel: Cell nuclei were costained with DAPI (blue). Right panel: the left panels were merged to the middle panels. Scale bar = 25 lm.

C Representative current traces recorded from HEK293 cells stably expressing either the wild-type NaV1.7 or the TAP-tagged NaV1.7 in response to depolarisation steps from 100 to 40 mV.

D Left panel: I(V) curves obtained in Nav1.7- and TAP-tagged NaV1.7-expressing cells using the same protocol as in (C), showing no significant difference in voltage of half-maximal activation (V1/2; 9.54 1.08 mV for NaV1.7, n = 10; and 8.12 1.07 mV for TAP-tagged NaV1.7, n = 14; P = 0.3765, Student’s t-test) and reversal

potential (Vrev; 61.73 3.35 mV for NaV1.7, n = 10; and 8.12 1.07 mV for TAP-tagged NaV1.7, n = 14; P = 0.9114, Student’s t-test). Right panel: Voltage dependence of fast inactivation, assessed by submitting the cells to a 500-ms prepulse from 130 to 10 mV prior to depolarisation at 10 mV. No significant

difference in the voltage of half inactivation was observed between the two cell lines (V1/2 of 64.58 1.32 mV for NaV1.7, n = 10; and 64.39 1.06 for TAP- tagged NaV1.7, n = 14; P = 0.9100, Student’s t-test). All data are mean SEM.

ª 2018 The Authors The EMBO Journal Vol 37 |No3 | 2018 429 Published online: January 15, 2018

The EMBO Journal NaV1.7-interacting proteins Alexandros H Kanellopoulos et al

Southern Blot – Na 1.7TAP-neo Mice C V WT A F1 F1

TAP Tag WT kb Na 1.7 kb V 20 20 WT WT 25 26 * 27 10 (13.8kb) 10 (13.8kb) 8 8 KI KI 6 (7.3kb) B 6 (6.6kb) I I 13.8kb Stu Stu 5.8kb HI HI I I 8.4kb 1 kb Bsp Stu

Bsp I / 3’ probe

Psi Stu Psi I / 5’ probe NaV1.7 allele 25 26 * 27 3’arm TAP 5’arm D Southern Blot – NaV1.7 Mice Targeting vector 3.4kb 5.8kb PmeI DTA neo

I kb

I kb I I

6.6kb Stu 7.3kb Stu Stu Stu 10 10 Targeted allele WT neo 8 8 (8.4kb) HI I HI Targeted allele I KI Psi

(Cre-LoxP Bsp Psi Bsp 4.5kb 6.1kb 6 recombination) 6 WT (6.1kb) (5.8kb) 5 5 * KI 4 Exon Stop Codon LoxP site TAP tag Probe Homologous arm 4 (4.5kb)

BspHI / 5’ probe PsiI / 3’ probe

E PCR F RT-PCR 170 bp 1.5 kb NaV1.7 27 25 26 Exon 27

TAP NaV1.7 27 TAP tag 27 2526 27 Tag Exon 27 411 bp 1.8 kb WT KI WT KI kb bp 3.0 500 KI (411 bp) 400 2.0 300 KI (1.8 kb) 200 WT (170 bp) 1.5 WT (1.5 kb) 100 1.0

Figure 2. Generation of TAP-tagged NaV1.7 knock-in mice.

A The location of TAP-tag in the NaV1.7 locus. A sequence encoding a TAP-tag peptide comprised of a HAT domain, TEV cleavage site and 3 FLAG-tags was inserted immediately prior to the stop codon (indicated as a star) at the extreme C-terminus of NaV1.7.

B Schematic diagrams of the targeting strategy. Grey boxes represent NaV1.7 exons (exon numbers are indicated on the box), black box represents TAP-tag, thick black lines represent homologous arms, and the small triangle box represents the single LoxP site, respectively. The positions of the external probes used for Southern blotting are indicated in the diagram. Neomycin (neo), DTA expression cassettes and restriction enzyme sites and expected fragment sizes for Southern blotting are also indicated.

C Southern blot analysis of genomic DNA from Founder 1 mice (TAP-tagged NaV1.7 carrying the neo cassette). Genomic DNA was digested with StuI and was then hybridised with either 50 or 30 external probe. Wild-type (WT) allele was detected as a 13.8 kb fragment using either 50 or 30 probes. Knock-in allele (KI) was detected as either a 6.6-kb (50 probe) or a 7.3-kb (30 probe) fragment.

D Southern blot analysis of TAP-tagged NaV1.7 allele after Cre recombination. Genomic DNA was digested either with BspHI or PsiI and was then hybridised with either 0 0 0 0 5 (BspHI) or 3 (PsiI) external probe. WT alleles were detected as 5.8 kb (5 probe) and 8.4 kb (3 probe) fragments, respectively. The neo-deleted TAP-tagged NaV1.7 alleles were detected as 4.5-kb (50 probe) and 6.1-kb (30 probe) fragments, respectively. TAP E Genotyping analysis by PCR. Representative result of the PCR screening of NaV1.7 mice showing the 411-bp band (KI allele) and the 170-bp band (WT allele). The location of primers used for PCR is indicated with black arrows. TAP F TAP-tagged NaV1.7 expression analysis with RT–PCR. Total RNA was isolated from DRG of NaV1.7 mice, and cDNA synthesis was primed using oligo-dT. PCR was performed with the primers indicated with black arrows. A 1.5-kb WT band and a 1.8-kb KI band were detected from either littermate WT control animals or TAP NaV1.7 KI mice, respectively.

430 The EMBO Journal Vol 37 |No3 | 2018 ª 2018 The Authors Published online: January 15, 2018

Alexandros H Kanellopoulos et al NaV1.7-interacting proteins The EMBO Journal

A B C D Rotarod von Frey Hargreaves Randall-Selio 80 1.0 8 240 0.8 200 60 6 160 0.6 40 4 120 0.4 Time (s) Time

Force (g) 80 20 Latency (s) 2 0.2 40

0 (g) Threshold Withdrawal 50% 0.0 0 0 WT KI WT KI WT KI WT KI

E Formalin test Formalin test 100 400 WT WT 80 KI KI

(sec) 300

60 200 40

Licking (sec) Licking 100 20 Licking/biting

0 0 0 5 10 15 20 25 30 35 40 45 50 Phase I Phase II Time (min)

Figure 3. Pain behaviour tests.

A Rotarod test showed no motor deficits in TAP-tagged NaV1.7 animals (n = 7, WT; n = 7, KI; P = 0.8619, Student’s t-test). B Responses to low-threshold mechanical stimulation by von Frey filaments were normal in the KI mice (P = 0.9237, Student’s t-test). C Hargreaves’ apparatus demonstrated identical latencies of response to thermal stimulation (n = 7, WT; n = 7, KI; P = 0.3038, Student’s t-test). D Acute mechanical pressure applied with the Randall–Selitto apparatus demonstrated identical behaviour in KI and WT mice (n = 7,WT;n = 7,KI;P = 0.7124,Student’s t-test). E Formalin test. Licking/biting response to acute peripheral inflammation induced by intraplantar injection of 5% formalin in hind-paw was recorded. Left panel, the time course of development of the response of KI mice (black squares) and WT littermate controls (white squares) showed similar response patterns (n = 10, WT; n = 7, KI; P = 0.2226, two-way ANOVA). Right panel, the early (0–10 min) and late (10–45 min) phases of the formalin response in KI and WT mice showed similar responses (P = 0.1690, first phase; P = 0.5017, second phase; Student’s t-test) between KI and WT mice. Data information: All data are mean SEM. neurons of the substantia nigra reticular region and the red nucleus different tissues with Western blot using an anti-HAT antibody. magnocellular region of the midbrain, and in neurons of the pontine TAP-tagged NaV1.7 bands were present in olfactory bulb, hypothala- nuclei located in the hindbrain (Fig 4J–L). In the spinal cord, FLAG- mus, spinal cord, sciatic nerve and DRG, but not detectable in tag expression was visible in the superficial layer of the dorsal horn. cortex, cerebellum, skin, lung, heart and pancreas (Fig 5E). Costaining of the spinal cord with isolectin B4 (IB4), a marker for the inner part of lamina II, showed that the TAP-tagged NaV1.7 was Optimisation of single-step- and tandem affinity purification of expressed in laminae I, II and III (Fig 4M–O). In the PNS, however, TAP-tagged NaV1.7 there was no positive FLAG-tag staining found in DRG, sciatic nerve or skin nerve terminals (data not shown). This could be because of The TAP-tag on NaV1.7 offered the possibility of two consecutive masking of the tag in the PNS, preventing the antibody binding. We affinity purifications, a single-step affinity purification (ss-AP; steps also examined the TAP-tagged NaV1.7 expression pattern in 1 and 2 in Fig 5A) and a tandem affinity purification (steps 1–4in

Figure 4. Immunohistochemistry of FLAG-tag expression in the central nervous system (CNS). ▸ A–D In the main olfactory bulb (MOB) (A), FLAG-tag expression (in green) is visible in the olfactory nerve layer (ONL) and in the glomerular layer (GL) in TAP-tagged

NaV1.7 knock-in mice (KI) but not in the littermate wild-type controls (WT). The white square box in (A) is shown with high magnification in (B). In the posterior olfactory bulb, staining is also evident in the (C) accessory olfactory bulb (AOB). Staining is absent in the (D) MOB and AOB of wild-type control mice. E, F FLAG-tag expression is present (E) in the medial habenula (MHb, arrow), the anterodorsal thalamic nucleus (AD, arrow), the laterodorsal thalamic nucleus (LD, dotted line) and (F) in the subfornical organ (SFO, arrow) located in the roof of the dorsal third ventricle. G–I FLAG-tag expression is present (G) in neurons of (H) the posterodorsal aspect of the medial amygdala (MePD, arrow, dotted line) and in (I) the hypothalamus in neurons of the arcuate nucleus (Arc, arrow, dotted line). J–L FLAG-tag expression is present (J) in neurons of the substantia nigra reticular part (SNR) and (K) the red nucleus magnocellular part (RMC) of the midbrain, and (L) in neurons of the pontine nuclei (Pn) located in the hindbrain. M, N The cross section of lumbar spinal cord (L4) is labelled with anti-FLAG (in red). FLAG-tag expression is present in laminae I, II and III in spinal cord of KI mice (N) but not in spinal cord of WT mice (M). O The cross sections of spinal cord of KI mice were costained with laminae II marker IB4 (in green). Data information: Sketches on the left illustrate the CNS regions and bregma levels (in mm) of the fluorescence images shown on the right. Scale bars: 500 lm (A, C, D, G); 250 lm (E, J, K, M–O); 100 lm (B, F, I, L); 50 lm (H). cp, cerebral peduncle; CTX, cortex; DM, dorsomedial hypothalamic nucleus; EPL, external plexiform layer; ME, median eminence; opt, optic tract; sm, stria medullaris; TH, thalamus; VMH, ventromedial hypothalamic nucleus; 3V, third ventricle.

ª 2018 The Authors The EMBO Journal Vol 37 |No3 | 2018 431 Published online: January 15, 2018

The EMBO Journal NaV1.7-interacting proteins Alexandros H Kanellopoulos et al

A CD

B

EF

GH

I

JKL

MN O I I II I II II III III III Spinal Cord L4

FLAG FLAG FLAG/IB4

Figure 4.

432 The EMBO Journal Vol 37 |No3 | 2018 ª 2018 The Authors Published online: January 15, 2018

Alexandros H Kanellopoulos et al NaV1.7-interacting proteins The EMBO Journal

Figure 5. Optimisation of single-step and tandem affinity purification, validation of identified protein–protein interactors and tissue expression pattern of TAP-tagged Na 1.7 in Na 1.7TAP mice. ▸ V V A Schematic illustrating the affinity purification (ss-AP and TAP) procedure using the tandem affinity tags separated with a TEV cleavage site.

B The proteins from DRG and olfactory bulbs were extracted in 1% CHAPS lysis buffer. After single-step and tandem affinity purification, TAP-tagged NaV1.7 was detected using Western blotting with anti-HAT antibody. C The proteins from different tissues including hypothalamus, sciatic nerve, spinal cord, olfactory bulbs and DRG from KI mice, and pooled tissues from WT mice were

extracted in 1% CHAPS lysis buffer. After single-step affinity purification, TAP-tagged NaV1.7 was detected using Western blotting with anti-HAT antibody.

D The interaction between TAP-tagged NaV1.7 and identified NaV1.7 protein–protein interactors including Scn3b, Syt2, Crmp2, Gprin1, Lat1 and Tmed10 was validated using a co-immunoprecipitation in vitro system. The expression vectors containing cDNA of validated genes were cloned and transfected into a HEK293 cell line

stably expressing TAP-tagged NaV1.7. After transfection, TAP-tagged NaV1.7 complexes were immunoprecipitated with anti-FLAG antibody, and the selected candidates were detected with their specific antibody using Western blotting. The results showed the expected sizes of Scn3b(32 kDa), Syt2 (44 kDa), Crmp2 (70 kDa), Gprin1 (two isoforms: 80 kDa and 110 kDa), Lat1 (57 kDa) and Tmed10 (21 kDa).

E Tissue expression pattern of TAP-tagged NaV1.7. The proteins were extracted from different tissues in both KI and WT littermate control mice and anti-FLAG used to detect TAP-tagged NaV1.7 using Western blotting. Anti b-tubulin was used as a loading control.

F The validation of selected NaV1.7 protein interactor candidates with Nav1.7 endogenous expressing DRG tissue. First, the proteins from DRG of TAP-tagged NaV1.7 mice were extracted in 1% CHAPS lysis buffer. NaV1.7 complexes were then immunoprecipitated by anti-FLAG M2 magnetic beads. Thirteen NaV1.7 interactor candidates including Scn3b(32 kDa), Syt2 (44 kDa), Crmp2 (70 kDa), Gprin1 (110 kDa), Lat1 (57 kDa), Tmed10 (21 kDa), Akap12 (191 kDa), Nfasc (138 kDa), Ntm (38 kDa), Kif5b(110 kDa), Ank3 (243 kDa) and Pebp1 (23 kDa) were detected with their specific antibodies using Western blotting.

G Co-immunoprecipitation of NaV1.7 with CaV2.2. Left panel shows negative Western blot results for pull-down of transiently transfected HA-tagged CaV2.2 from TAP-

tagged NaV1.7 complex (HAT antibody for detection) in TAP-tagged NaV1.7 HEK293 stable cell line. Right panel shows control blot from whole-cell lysate of HA-tagged CaV2.2 and TAP-tagged NaV1.7. Source data are available online for this figure.

Fig 5A). CHAPS and DOC lysis buffers were evaluated to solubilise components (Fig 6A). These proteins were further classified into 22

TAP-tagged NaV1.7 and its protein complex from tissues, for exam- groups based on their function, including 12 membrane-trafficking ple DRG, spinal cord, olfactory bulbs and hypothalamus, in the co- proteins, 23 enzyme modulators and four transcription factors IP system. CHAPS is a non-denaturating zwitterionic detergent, (Fig 6B). commonly used to extract membrane proteins in their native confor-

mation. In comparison with strong anionic detergents like SDS, Validation of TAP-tagged NaV1.7-interacting proteins using co-IP CHAPS preserves protein–protein interactions and is compatible

with downstream applications such as mass spectrometry. The The physical interactions between NaV1.7 and interacting protein

result showed that TAP-tagged NaV1.7 from DRG and olfactory candidates were assessed using co-IP with DRG tissue extracts

bulbs was clearly solubilised and precipitated by both purifications from TAP-tagged NaV1.7 mice. A number of candidates of interest, —ss-AP and tandem affinity purification in 1% CHAPS buffer such as Scn3b, Syt2, Lat1, Tmed10, Gprin1, Crmp2, isoform 2 of (Fig 5B). Also, the result from ss-AP showed that TAP-tagged A-kinase anchor protein 12 (Akap12), neurofascin (Nfasc), neuro-

NaV1.7 could be immunoprecipitated from hypothalamus, sciatic trimin (Ntm), kinesin-1 heavy chain (Kif5b), ankyrin-3 (Ank3) and nerve, spinal cord, olfactory bulb and DRG of KI mice, but not from phosphatidylethanolamine-binding protein 1 (Pebp1), were chosen

these tissues of WT control mice (Fig 5C). However, the DOC lysis from the NaV1.7-associated protein list previously identified by

buffer, which was used to investigate the TAP-tagged PSD-95 MS (Table 1 and Table EV1). After ss-AP, the NaV1.7 complexes protein complex (Fernandez et al, 2009), did not solubilise TAP- were separated by SDS–PAGE and the protein interactors of

tagged NaV1.7 from mouse tissue (data not shown). NaV1.7 were detected by Western blotting. Our results show that all 12 candidates were detected by their specific antibodies

Identification of TAP-tagged NaV1.7-associated complexes (Fig 5F). by AP-MS We also validated six candidates (Scn3b, Syt2, Crmp2, Gprin1, Lat1 and Tmed10) using co-IP in an in vitro system by co-expres-

We next identified the components of NaV1.7 complexes using ss- sing the candidates in a TAP-tagged NaV1.7 stable cell line. The AP followed by Liquid chromatography–tandem mass spectrometry mammalian expression vectors carrying cDNAs of these candi-

(LC-MS/MS). Briefly, the TAP-tagged NaV1.7 complexes were dates were transfected into a TAP-tagged NaV1.7 stable HEK293 extracted from DRG, spinal cord, olfactory bulb and hypothalamus cell line; 48 h of post-transfection, the proteins in the transfected

using ss-AP (see Materials and Methods). In total, 189,606 acquired cells were extracted. The TAP-tagged NaV1.7 multiprotein spectra from 12 samples, in which each group (KI and WT) contains complexes were then immunoprecipitated with anti-FLAG anti- six biological replicate samples and each sample was from one body and analysed with Western blot using different specific anti- mouse, were used for protein identification; 1,252 proteins were bodies against those selected candidates. The Western blot identified with a calculated 0.96% false discovery rate (FDR); 267 showed that all six candidates were detected with the expected proteins (Table EV1) met those criteria and were shortlisted sizes on the blot (Fig 5D), confirming that these candidates are

based on the criteria described in Materials and Methods. The contained in the NaV1.7 complex. In summary, all these potential TAP proteins only appearing in NaV1.7 mice and the representatively interacting proteins selected for further validation were confirmed selected proteins are listed in Table 1. PANTHER cellular compo- as interactors by co-IP, giving confidence in the Nav1.7TAP mass nent analysis of these 267 proteins revealed eight different cellular spectrometry list.

ª 2018 The Authors The EMBO Journal Vol 37 |No3 | 2018 433 Published online: January 15, 2018

The EMBO Journal NaV1.7-interacting proteins Alexandros H Kanellopoulos et al

A Step 1: Flag Beads Capture Step 2: TEV Protease Cleavage Step 3: Ni2+ Beads Capture Step 4: Elution

HAT HAT HAT Nav1.7 Nav1.7 Nav1.7 Nav1.7

FLAG Beads Ni Beads Ni Beads FLAG Beads

B C Crude Lysate ss-AP TAP

kDa WT KI WT KI WT KI kDa

250 250

D E

kDa kDa 250 HAT 250 KI β-Tubulin 50 72 HAT 250 52 WT β-Tubulin 30 50

F kDa

250

72

52

30

G kDa HAT HA HAT HA

250

Figure 5.

Functional characterisation of Crmp2 with Crmp2 acting directly as a transporter for NaV1.7 (Dustrude et al, 2013). Next, we sought to investigate the effect of lacosa- We confirmed that Tap-tagged Nav1.7 binds directly to Crmp2, mide on sodium currents. In cells not transfected with Crmp2,

the presumed target of the anti-epileptic and analgesic drug laco- NaV1.7 currents displayed very small changes in current density samide, based on binding studies (Wilson & Khanna, 2015). We under our recording conditions following prolonged 5-h exposure sought to evaluate the possible electrophysiological consequences to lacosamide (Fig 7A and B). Interestingly, following 5-h incuba-

of Crmp2 binding to NaV1.7 and in addition to understand the tion with lacosamide in Crmp2-transfected cells, a complete rever-

relevance of Crmp2 to the action of lacosamide. Transfection of sal in the increase in NaV1.7 current density provoked by Crmp2

Crmp2 into a NaV1.7 stable HEK293 cell line revealed a nearly was observed (Fig 7A and D). This shows an interaction between

twofold increase in sodium current density (Fig 7A–C), consistent NaV1.7 and Crmp2 and demonstrates that actions of lacosamide

434 The EMBO Journal Vol 37 |No3 | 2018 ª 2018 The Authors Published online: January 15, 2018

Alexandros H Kanellopoulos et al NaV1.7-interacting proteins The EMBO Journal

Table 1. Identified NaV1.7-associated proteins (only appearing in KI group + selected candidates). Normalised spectral index quantity

UniProt KI WT Ratio Gene symbol Protein name Acc (Mean) (Mean) (KI/WT) Proteins appearing in KI group only Scn3b Sodium channel subunit beta-3 Q8BHK22.64E06 0.00E+00 KI only Bola2 BolA-like protein 2 Q8BGS22.28E06 0.00E+00 KI only Psma7 Proteasome subunit alpha type-7 Q9Z2U01.77E06 0.00E+00 KI only Psma1 Proteasome subunit alpha type-1 Q9R1P41.70E06 0.00E+00 KI only Homer2 Isoform 2 of Homer protein homolog 2 Q9QWW1-21.22E06 0.00E+00 KI only Fga Isoform 2 of Fibrinogen alpha chain E9PV24-21.06E06 0.00E+00 KI only Ndufv3 NADH dehydrogenase [ubiquinone] flavoprotein 3, mitochondrial Q8BK30 9.48E07 0.00E+00 KI only Psma5 Proteasome subunit alpha type-5 Q9Z2U19.22E07 0.00E+00 KI only Erh Enhancer of rudimentary homolog P84089 9.12E07 0.00E+00 KI only Psmb2 Proteasome subunit beta type-2 Q9R1P38.92E07 0.00E+00 KI only Rp2 Isoform 2 of Protein XRP2 Q9EPK2-28.67E07 0.00E+00 KI only Pebp1 Phosphatidylethanolamine-binding protein 1 P70296 8.32E07 0.00E+00 KI only Plekhb1 Isoform 2 of Pleckstrin homology domain-containing Q9QYE9-28.18E07 0.00E+00 KI only family B member 1 Omp Olfactory marker protein Q64288 6.62E07 0.00E+00 KI only Ndufb10 NADH dehydrogenase [ubiquinone] 1 beta subcomplex subunit 10 Q9DCS96.24E07 0.00E+00 KI only Fabp7 Fatty acid-binding protein, brain P51880 5.87E07 0.00E+00 KI only Apoo Apolipoprotein O Q9DCZ45.53E07 0.00E+00 KI only Psmb5 Proteasome subunit beta type-5 O55234 4.79E07 0.00E+00 KI only Psmd12 26S proteasome non-ATPase regulatory subunit 12 Q9D8W54.47E07 0.00E+00 KI only Mlst8 Target of rapamycin complex subunit LST8 Q9DCJ14.22E07 0.00E+00 KI only Flna Filamin-A Q8BTM84.00E07 0.00E+00 KI only Psma4 Proteasome subunit alpha type-4 Q9R1P03.79E07 0.00E+00 KI only Oxct1 Succinyl-CoA:3-ketoacid coenzyme A transferase 1, mitochondrial Q9D0K23.72E07 0.00E+00 KI only Psmb4 Proteasome subunit beta type-4 P99026 3.54E07 0.00E+00 KI only Tusc5 Tumour suppressor candidate 5 homolog Q8C838 3.22E07 0.00E+00 KI only Map2 Microtubule-associated protein (Fragment) G3UZJ22.41E07 0.00E+00 KI only Psmd4 Isoform Rpn10Bof26S proteasome non-ATPase regulatory subunit 4 O35226-22.13E07 0.00E+00 KI only Rab1A Ras-related protein Rab-1AP62821 1.90E07 0.00E+00 KI only Calb2 Calretinin Q08331 1.83E07 0.00E+00 KI only Actr2 Actin-related protein 2 P61161 1.34E07 0.00E+00 KI only Akap12 Isoform 2 of A-kinase anchor protein 12 Q9WTQ5-21.23E07 0.00E+00 KI only Gpx4 Isoform Cytoplasmic of Phospholipid hydroperoxide O70325-21.15E07 0.00E+00 KI only glutathione oxidase Ahsa1 Activator of 90-kDa heat-shock protein ATPase homolog 1 Q8BK64 1.12E07 0.00E+00 KI only Htatsf1 HIV Tat-specific factor 1 homolog Q8BGC09.15E08 0.00E+00 KI only Sart3 Squamous cell carcinoma antigen recognised by T cells 3 Q9JLI88.80E08 0.00E+00 KI only Wdr7 WD repeat-containing protein 7 Q920I98.24E08 0.00E+00 KI only Rbm10 Isoform 3 of RNA-binding protein 10 Q99KG3-37.90E08 0.00E+00 KI only Tkt Transketolase P40142 7.83E08 0.00E+00 KI only Tmed10 Transmembrane emp24 domain-containing protein 10 Q9D1D47.75E08 0.00E+00 KI only Ntm Neurotrimin Q99PJ07.69E08 0.00E+00 KI only Lnp Protein lunapark Q7TQ95 7.52E08 0.00E+00 KI only

ª 2018 The Authors The EMBO Journal Vol 37 |No3 | 2018 435 Published online: January 15, 2018

The EMBO Journal NaV1.7-interacting proteins Alexandros H Kanellopoulos et al

Table 1 (continued) Normalised spectral index quantity

UniProt KI WT Ratio Gene symbol Protein name Acc (Mean) (Mean) (KI/WT) Ddb1 DNA damage-binding protein 1 Q3U1J47.32E08 0.00E+00 KI only Mpp2 Isoform 2 of MAGUK p55 subfamily member 2 Q9WV34-26.92E08 0.00E+00 KI only Camsap3 Isoform 2 of calmodulin-regulated spectrin- Q80VC9-26.56E08 0.00E+00 KI only associated protein 3 Ppp1r9a Protein Ppp1r9aH3BJD06.25E08 0.00E+00 KI only Pccb Propionyl-CoA carboxylase beta chain, mitochondrial Q99MN96.16E08 0.00E+00 KI only Des Desmin P31001 5.91E08 0.00E+00 KI only Cdh2 Cadherin-2 P15116 5.88E08 0.00E+00 KI only Agpat3 1-acyl-sn-glycerol-3-phosphate acyltransferase gamma Q9D517 5.50E08 0.00E+00 KI only Ogt Isoform 2 of UDP-N-acetylglucosamine Q8CGY8-25.42E08 0.00E+00 KI only Ahnak Protein Ahnak E9Q616 5.34E008 .00E+00 KI only Nfasc Neurofascin Q810U34.96E08 0.00E+00 KI only Mccc2 Methylcrotonoyl-CoA carboxylase beta chain, mitochondrial Q3ULD54.63E08 0.00E+00 KI only Mpdz Isoform 2 of Multiple PDZ domain protein Q8VBX6-23.66E08 0.00E+00 KI only Tln1 Talin-1 P26039 3.32E08 0.00E+00 KI only Dsp Desmoplakin E9Q557 2.17E08 0.00E+00 KI only Slc4a8 Isoform 2 of Electroneutral sodium bicarbonate exchanger 1 Q8JZR6-21.82E08 0.00E+00 KI only Slc7a14 Probable cationic amino acid transporter Q8BXR11.07E08 0.00E+00 KI only Sgk223 Tyrosine-protein kinase SgK223 Q571I47.46E09 0.00E+00 KI only Other selected candidates for validation Slc7a5/Lat1 Large neutral amino acids transporter small subunit 1 Q9Z127 1.66E07 3.11E08 5.34 Kif5b Kinesin-1 heavy chain Q61768 4.04E07 1.00E07 4.04 Ank3 Ankyrin-3 (Fragment) S4R2K93.06E07 8.73E08 3.51 Dpysl2/Crmp2 Dihydropyrimidinase-related protein 2 O08553 7.58E06 2.26E06 3.35 Gprin1 G protein-regulated inducer of neurite outgrowth 1 Q3UNH43.94E07 2.05E07 1.92 Syt2 Synaptotagmin-2 P46097 3.81E06 2.46E06 1.55

MGI approved gene symbols and protein names, and UniProt accession numbers are shown. Average of spectral counts from TAP-tagged Nav1.7 knock-in mice (KI) and littermate wild-type control mice (WT) and the ratios from KI group and WT group are displayed.

on Nav1.7 function are predominantly mediated through Crmp2 knockout mice (Nassar et al, 2004). However, there was no obvious binding. alteration of CGRP or SP levels in lamina I and II in the spinal cord

It has been shown that the activity of N-type voltage-gated in NaV1.7 knockout mice compared to wild-type littermates

calcium channels (CaV2.2) in relation to neurotransmitter release (Fig EV2). Thus, levels of neurotransmitters are unaffected by

from presynaptic terminals in pain pathways is regulated by its inhibiting release in NaV1.7 null mutants. protein interactor Crmp2 (Chi et al, 2009; Brittain et al, 2011).

Our study now demonstrates that Crmp2 is also a direct protein Interactions between Nav1.7 and opioid receptor signalling

interactor of Nav1.7. Therefore, we tested the possibility that

Nav1.7 is involved in neurotransmitter release linked to Cav2.2 Recently a massive potentiation of opioid signalling has been

through Crmp2 using co-IP in an in vitro system in HEK described in sensory neurons of NaV1.7 null mutant mice (Isensee

cells. However, the results show that Cav2.2 was not immuno- et al, 2017). This effect is specific for NaV1.7 as it is not replicated

precipitated together with Nav1.7 (Fig 5G), suggesting that in NaV1.8 null mutant mice. We found that Gprin1, a l-opioid

CaV2.2 and NaV1.7 may regulate neurotransmitter release receptor-binding protein (Ge et al, 2009), is also associated with

independently. NaV1.7 (Table 1). Furthermore, the protein interaction between

We further investigated the function of NaV1.7 in relation to Gprin1 and NaV1.7 was confirmed with co-IP in vitro (Fig 5D).

presynaptic neurotransmitter release using immunohistochemistry This suggests close proximity between the sodium channel NaV1.7 to identify distribution changes in neurotransmitter CGRP and and opioid receptors whose efficacy is known to be regulated by

substance P (SP) in the dorsal horn of the spinal cord in NaV1.7 sodium.

436 The EMBO Journal Vol 37 |No3 | 2018 ª 2018 The Authors Published online: January 15, 2018

Alexandros H Kanellopoulos et al NaV1.7-interacting proteins The EMBO Journal

A Cellular Component B Protein class 1% Hydrolase 41 Cytoskeletal protein 34 16% Transporter 26 1% Enzyme modulator 23 Nucleic acid binding 21 33% Chaperone 16 Transferase 15 Oxidoreductase 13 22% Membrane traffic protein 12 Transfer/carrier protein 10 Calcium-binding proteins 10 Structural protein 7 2% Ligase 7 11% 14% Cell adhesion molecule 7 Lyase 6 Transcription factor 4 Cell junction Signaling molecule 4 Cell part Receptor 4 Extracellular matrix Defense/immunology protein 4 Macromolecular complex Isomerase 3 Membrane Cell junction protein 2 Organelle Extracellular matrix protein 1 Synapse 0 1020304050 Unknown Number of protein

Figure 6. Analysis of NaV1.7 complex proteins.

A Cellular localisation of identified NaV1.7-interacting proteins. B Protein class of identified NaV1.7-interacting proteins categorised using PANTHER Classification System.

Discussion single-step FLAG approach, Chen and colleagues defined specific novel interactors for the catalytic subunit of PP4, which they had Experimental evidence has shown that protein–protein interactions not previously observed with TAP-MS (Chen & Gingras, 2007). play a key role in trafficking, distribution, modulation and stability of Thus, ss-AP followed by sensitive LC-MS/MS analysis was applied in ion channels (Shao et al, 2009; Catterall, 2010; Leterrier et al,2010; this study. In fact, many dynamic modulator proteins were identified

Bao, 2015; Chen-Izu et al, 2015; Laedermann et al,2015).Here,we to interact with NaV1.7 in this study, such as calmodulin (Calm1;

mapped the protein interaction network of NaV1.7 using an AP-MS Table EV1), which was found to bind to the C-terminus of other

proteomic approach with an epitope-tagged NaV1.7 knock-in mouse VGSCs NaV1.4 and NaV1.6, thereby regulating channel function line. This is the first report to define an ion channels’ macromolecular (Herzog et al, 2003). Apart from the sodium channel b3 subunit and

complex using an epitope-tagged gene-targeted mouse. the known NaV1.7 protein interactor Crmp2, a broad range of impor- AP-MS requires specific high-affinity antibodies against the target tant novel interactors that belong to different protein classes, such as proteins of interest (Wildburger et al, 2015). However, binding may cytoskeletal/structural/cell-adhesion proteins and vesicular/traf- compete with normal protein–protein interactions. To overcome ficking/transport proteins (Fig 6B), have been identified in this study. these limitations, epitope tags on target proteins were introduced Four proteins, Crmp2, Nedd4-2, FGF13 and Pdzd2, have previously

into the AP-MS system. In the last decade, single-step and tandem been reported as NaV1.7 protein interactors (Sheets et al, 2006; Shao affinity purification have been widely applied in protein–protein et al,2009;Hoet al, 2012; Yang et al,2017).Laedermannet al

interaction studies (Fernandez et al, 2009; Wildburger et al, 2015). (2013) showed that the E3 ubiquitin ligase Nedd4-2 regulates NaV1.7 In contrast to ss-AP, TAP produces lower background and less by ubiquitinylation in the pathogenesis of neuropathic pain. Shao contamination. However, due to its longer experimental washing et al (2009) demonstrated that Pdzd2 binds to the second intracellu-

procedures and two-step purification, TAP coupled with MS analysis lar loop of NaV1.7 by a GST pull-down assay in an in vitro system. may not be sufficiently sensitive to detect transient and dynamic Recently, Zhang’s group revealed that FGF13 regulates heat nocicep-

protein–protein interactions. In recent years, along with newly tion by interacting with NaV1.7 (Yang et al,2017).Wedidnotfind

developed highly sensitive mass spectrometer techniques and the previously reported NaV1.7 interactors Nedd4-2, FGF13 or Pdzd2.

powerful quantitative proteomics analysis methods, ss-AP was This may be because Nedd4-2 and FGF13 only bind to NaV1.7 in employed to identify both transient and stable protein–protein inter- neuropathic pain conditions and in heat nociception, respectively, actors (Oeffinger, 2012; Keilhauer et al, 2015). For example, using a and Pdzd2 shows strong binding in vitro but not in vivo.

ª 2018 The Authors The EMBO Journal Vol 37 |No3 | 2018 437 Published online: January 15, 2018

The EMBO Journal NaV1.7-interacting proteins Alexandros H Kanellopoulos et al

A NaV1.7 NaV1.7 + 100 μM LCM

NaV1.7 + Crmp2 NaV1.7 + Crmp2 + 100 μM LCM

40 mv

1nA -80 mv 100 ms

BC 10 10

0 0

-10 -10

-20 -20 I (pA/pF) -30 I (pA/pF) -30

NaV1.7 NaV1.7 -40 + 100 μM LCM -40 + Crmp2 Control Control -50 -50 -50 0 -50 0 Voltage (mV) Voltage (mV)

D 10

0

-10

/pF) -20

I (pA -30 Nav1.7 + Crmp2 -40 + 100 μM LCM Control -50 -50 0 Voltage (mV)

Figure 7. Electrophysiological characterisation of NaV1.7 in HEK293 cells following transfection with Crmp2 and incubation with lacosamide (LCM).

A Representative raw current traces of NaV1.7 stably expressed in HEK293 cells in response to the activation pulse protocol shown. Each trace shows a different condition: + Crmp2 transfection and + incubation with 100 lM LCM.

B IV plot of NaV1.7 current density in the absence and presence of 100 lM LCM. Compared with NaV1.7 basal currents (n = 10), LCM incubation had no significant effect on sodium channel density (n = 12, P = 0.402).

C IV plot of NaV1.7 in HEK293 cells in the presence and absence of Crmp2 transfection. Compared to NaV1.7 basal currents, Crmp2 transfection caused a significant

increase in NaV1.7 current density (n = 16, P = 0.0041). D IV plot showing current density of NaV1.7 following transfection of Crmp2 and incubation with 100 lM LCM. Incubation with LCM reversed the Crmp2-mediated current increase (n = 10, P = 0.0442). Data information: Data were analysed using one-way ANOVA with Tukey’s post hoc test. All data are mean SEM.

Scn3b and Crmp2 have previously been proposed to associate expressing NaV1.7 HEK293 cells and that this upregulation can be

with NaV1.7 (Ho et al, 2012; Dustrude et al, 2013). We demon- reversed by applying lacosamide. Previous studies reported that no

strated the physical interaction between these proteins and NaV1.7 change in NaV1.7 currents with Crmp2 overexpression in CAD cells by co-IP. Furthermore, we demonstrated that transient overexpres- (Wang et al, 2010; Dustrude et al, 2013). This could be attributed

sion of Crmp2 can upregulate NaV1.7 current density in stably perhaps to the different species’ Crmp2 or the different transfection

438 The EMBO Journal Vol 37 |No3 | 2018 ª 2018 The Authors Published online: January 15, 2018

Alexandros H Kanellopoulos et al NaV1.7-interacting proteins The EMBO Journal

conditions. Taken together, FLAG ss-AP coupled with quantitative NaV1.7 has also been linked to opioid peptide expression, and

MS seems to be a powerful and reliable tool for investigating protein enhanced activity of opioid receptors is found in the NaV1.7 null interactions of membrane ion channels. mutant mouse (Minett et al, 2015). Interestingly, Gprin1, which is Using co-IP and a co-expression in an in vitro system, we also con- known to interact with opioid receptors as well as other GPCRs, was

firmed the direct physical interaction between NaV1.7 and novel found to co-immunoprecipitate with NaV1.7. This suggests that protein interactors of interest including Syt2, Lat1, Tmed10 and GPCR sodium channel interactions could add another level of regu-

Gprin1. Synaptotagmin is a synaptic vesicle membrane protein that latory activity to the expression of NaV1.7. Intriguingly, opioid functions as a calcium sensor for neurotransmission (Chapman, 2002). receptors are well known to be regulated by sodium (Ott et al, 1988;

Sampo et al (2000) showed a direct physical binding of synaptotag- Fenalti et al, 2014). Deleting Nav1.8 had no effect on l-opioid recep-

min-1 with NaV1.2 at a site, which is highly conserved across all volt- tor efficacy measured with fentanyl, whilst Nav1.7 deletion potenti- age-gated sodium channels suggesting that the synaptotagmin family ated opioid action substantially (preprint: Kanellopoulos et al,

can associate with other VGSCs. NaV1.7 found at presynaptic terminals 2017). Thus, the proximity of Nav1.7 and l-opioid receptors medi- appears to be involved in neurotransmitter release (Weiss et al,2011; ated by Gprin1 may contribute to this regulation.

Black et al, 2012; Minett et al, 2012). Our results revealed the physical More recently, Branco and colleagues reported that NaV1.7 in

interaction between Syt2 and NaV1.7. Interestingly, other synaptic hypothalamic neurons plays an important role in body weight

proteins also coprecipitated with Nav1.7 including SNARE complex control (Branco et al, 2016). We found that NaV1.7 was not only

protein syntaxin-12 (Table EV1). These data suggest that NaV1.7 may present in the arcuate nucleus but also in other regions of the brain regulate neurotransmitter release through Syt2 and syntaxin-12 in the such as the medial amygdala, medial habenula, anterodorsal thala- peripheral and central terminals of the spinal cord. mic nucleus, laterodorsal thalamic nucleus, and in the subfornical Gabapentin was developed to treat epilepsy, but it is now used to organ, substantia nigra reticular part and the red nucleus magnocel- treat various forms of chronic pain. However, the analgesic mecha- lular part of the midbrain, and in neurons of the pontine nuclei

nisms of gabapentin are not entirely clear, but involve the inhibition located in the hindbrain. NaV1.7 thus has other functions in the CNS

of the voltage-gated calcium channel CaV2.2 by blocking the traf- that remain to be elucidated.

ficking of a2d-1- and a2d-1-mediated trafficking of the CaV2.2 chan- Overall, the present findings provide new insights into the inter-

nel complex to reduce channel activity in DRG neurons (Hendrich actome of NaV1.7 for advancing our understanding of NaV1.7 func- et al, 2008; Cassidy et al, 2014). As the transporter of gabapentin, tion. Our data also show that the ss-AP-coupled LC-MS/MS is a

Lat1 has been identified as a NaV1.7 protein interactor, and it is sensitive, reliable and high-throughput approach to identify protein– conceivable that Lat1 may also be involved in the regulation of protein interactors for membrane ion channels, using epitope-tagged

NaV1.7 channel function. However, future experiments are required gene-targeted mice.

to determine the functional significance of the Nav1.7-Lat1 interac- tion in this regard. As a member of the p24 family, Tmed10 was selected due to its well-characterised properties as a protein traf- Materials and Methods ficking regulator. Previous studies have highlighted that Tmed10

plays an important role as a cargo receptor in the trafficking of vari- Generation of a TAP-tagged NaV1.7-expressing stable HEK293 ous membrane proteins. For example, Tmed10 has been observed cell line to modulate the transport and trafficking of amyloid-b precursor

protein (Vetrivel et al, 2007), endogenous glycosylphosphatidylinos- A HEK293 cell line stably expressing TAP-tagged NaV1.7 was estab- itol (GPI)-anchored proteins CD59 and folate receptor alpha (Bon- lished as previously described (Koenig et al, 2015). Briefly, a sequence non et al, 2010), and several G protein-coupled receptors (GPCRs) encoding a TAP-tag (peptide: SRK DHL IHN VHK EEH AHA HNK IEN

(Luo et al, 2011). Tmed10 was confirmed as a NaV1.7 protein inter- LYF QGE LPT AAD YKD HDG DYK DHD IDY KDD DDK) was inserted

actor in this study, suggesting Tmed10 may regulate NaV1.7 traf- immediately prior to the stop codon of NaV1.7 in the SCN9A mamma- ficking. The underlying mechanisms need to be further investigated. lian expression construct FLB (Cox et al, 2006). The TAP-tag at the

VGSCs are known to exist in macromolecular complexes extreme C-terminus of NaV1.7 comprises a HAT domain and 3 FLAG- (Meadows & Isom, 2005). The b subunits are members of the tags, enabling immunodetection with either anti-HAT or anti-FLAG

immunoglobulin (Ig) domain family of cell-adhesion molecules antibodies. The function and expression of TAP-tagged NaV1.7 in this (CAM). As well as sodium channel a subunits, the b subunits also HEK293 cell line were characterised with both immunocytochemistry bind to a variety of cell-adhesion molecules such as neurofascin, and electrophysiological patch clamp analysis. contactin, tenascins and NrCAMs (Srinivasan et al, 1998; Ratcliffe TAP et al, 2001; McEwen & Isom, 2004; Cusdin et al, 2008; Namadurai Generation of NaV1.7 knock-in mice et al, 2015). In our data set (Table EV1), the sodium channel b3 subunit and some CAMs, such as Ncam1 and neurofascin, have The gene targeting vector was generated using a BAC homologous

been found to associate with the NaV1.7 a subunit. The crystal recombineering-based method (Liu et al, 2003). Four steps were structure of the human b3 subunit has been solved recently. The b3 involved in this procedure (Fig EV1). Step 1, two short homology subunit Ig-domain assembles as a trimer in the crystal asymmetric arms (HA) HA3 and HA4 corresponding to 509-bp and 589-bp

unit (Namadurai et al, 2014). This raises the possibility that trimeric sequences within intron 26 and after exon 27 of NaV1.7,

b3 subunits binding to NaV1.7 a subunit(s) form a large complex respectively, were amplified by PCR using a BAC bMQ277g11 together with other sodium channels, as well as with CAMs and (Source Bioscience, Cambridge, UK) DNA as a template, and then cytoskeletal proteins in the plasma membrane. inserted into a retrieval vector pTargeter (Fernandez et al, 2009; gift

ª 2018 The Authors The EMBO Journal Vol 37 |No3 | 2018 439 Published online: January 15, 2018

The EMBO Journal NaV1.7-interacting proteins Alexandros H Kanellopoulos et al

from Dr. Seth GN Grant) by subcloning. Step 2, a 9.1-kb genomic E25-26-F (forward)—CCGAGGCCAGGGAACAAATTCC DNA fragment (3.4 kb plus 5.8 kb) was retrieved through homol- 30UTR-R (reverse)—GCCTGCGAAGGTGACTCACTCGTG

ogous recombineering by transforming the KpnI-linearised pTar- The wild-type and TAP-tagged NaV1.7 alleles gave a 1,521-bp geter-HA3-HA4 vector into EL250 E. coli cells containing BAC band and a 1,723-bp band, respectively. bMQ277g11. Step 3, short homology arms HA1 and HA2 corre-

sponding to 550 bp (before stop codon of NaV1.7) and 509 bp Immunocytochemistry

(starting from stop codon of NaV1.7), respectively, were amplified

by PCR, and then cloned into pneoflox vector (Fernandez et al, TAP-tagged NaV1.7-HEK293 cells and their parental HEK293 cells 2009; gift from Dr. Seth GN Grant) containing the TAP-tag, leav- were plated on poly-D-lysine coated coverslips in 24-well plates

ing inbetween the TAP-tag sequence, 2 LoxP sites, PGK and EM7 and cultured at 37°C/5% CO2 in DMEM supplemented with 10% promoters, the G418r gene and a SV40 polyadenylation site. Step foetal bovine serum (Life Technologies), 50 U/ml penicillin, 4, the cassette flanked by two homology arms (HA1 and HA2) 50 lg/ml streptomycin and 0.2 mg/ml G418 (only for TAP-tagged

was excised by XhoI and BglII digestion and transformed into NaV1.7-HEK293 cells); 24 h later, cells were fixed in cooled recombination-competent EL250 cells containing the pTargeter- methanol at 20°C for 10 min and then permeabilised with cooled HA3-HA4 plasmid. Then, the TAP tag cassette was inserted into acetone at 20°C for 1 min. After three washes with 1× PBS, the the pTargeter-HA3-HA4 vector by recombination in EL250 E. coli cells were incubated with blocking buffer containing 1× PBS, 0.3% cells. The correct recombination and insertion of the targeting Triton X-100 and 10% goat serum at room temperature for cassette were confirmed by restriction mapping and DNA 30 min. Then, the fixed cells were incubated with anti-FLAG anti- sequencing. The complete gene targeting vector containing the 50- body (1:500 in blocking buffer, F1804, Sigma) at 4°C overnight. 0 × end homologous NaV1.7 sequence of 3.4 kb and a 3 -end homol- After 3 washes with 1 PBS, cells were incubated with secondary ogy arm of 5.8 kb was linearised with PmeI digestion for ES cell antibody goat anti-mouse IgG conjugated with Alexa Fluor 488 electroporation. All the homology arms HA1, HA2, HA3 and HA4 (A11017, Invitrogen) at room temperature for 2 h. Then, coverslips were amplified with NEB Phusion PCR Kit using bMQ277g111 were washed 3 times in 1× PBS and cells were mounted with BAC clone DNA as a template. Primers used to create the recom- VECTASHIELD HardSet Antifade Mounting Medium containing bination arms included: DAPI (H-1400, Vectorlabs) and visualised using a fluorescence HA1XhoIF (HA1, forward)— acactcgagAGCCAAACAAAGTCCAGCT microscope (Leica). HA1XbaIR (HA1, reverse)—tgttctagaTTTCCTGCTTTCGTCTTTCTC HA2Acc65IF (HA2, forward)—tgaggtacctagAGCTTCGGTTTTGAT Immunohistochemistry ACACT HA2BglIIR (HA2, reverse)—gatagatctTTGATTTTGATGCAATGTAGGA Following anaesthesia, mice were transcardially perfused with PBS, HA3SpeIF (HA3, forward)—ctcactagtCTCTTCATACCCAACATGCCTA followed by 1% (for brain) or 4% (for spinal cord) paraformalde- HA3KpnIR (HA3, reverse)—aatggtaccGGATGGTCTGGGACTCCATA hyde in PBS. Brains and spinal cord were dissected and incubated in HA4KpnIF (HA4, forward)—gaaggtaccGCTAAGGGGTCCCAAATTGT fixative for 4 h at 4°C, followed by 30% sucrose in PBS for 2 days at HA4PmeIR (HA4, reverse)—tcagtttaaacGGGATGGGAGATTACGAGGT 4°C. Tissue was embedded in O.C.T. (Tissue-Tek) and snap-frozen

To generate TAP-tagged NaV1.7 mice, the linearised targeting in a dry ice/2-methylbutane bath. Brain coronal and spinal cord vector was transfected into 129/Sv ES cells. Cells resistant to G418 cross cryosections (20 lm) were collected on glass slides (Super- were selected by culturing for 9 days. Recombined ES cell clones frost Plus, Polyscience) and stored at 80°C until further process- were identified using Southern blot-based screening. Three clones ing. For FLAG-tag immunohistochemistry in the brain, sections that were confirmed to be correct using Southern blot were injected were incubated in blocking solution (4% horse serum, 0.3% Triton into C57BL/6 blastocysts at the Transgenic Mouse Core facility of X-100 in PBS) for 1 h at room temperature, followed by incubation the Institute of Child Heath (ICH). The chimeric animals were in mouse anti-FLAG antibody (F-1804, Sigma) diluted 1:200 in crossed to C57BL/6, and the germline transmission was confirmed blocking solution for 2 days at 4°C. Following three washes in PBS, by Southern blot. The neomycin cassette was removed by crossing bound antibody was visualised using either an Alexa 488- or 594- with global Cre mice. The correct removal of the neomycin cassette conjugated goat anti-mouse secondary antibody (1:800, Invitrogen). and TAP-tag insertion was confirmed by Southern blot, genotyping To facilitate the identification of brain regions and the correspond- (PCR) and RT–PCR. ing bregma levels, nuclei were counterstained with Hoechst 33342 The genomic DNA was extracted from ear punches and geno- (1:10,000, Invitrogen). Sections were mounted with DAKO fluores- typing analysis was achieved using a standard PCR. Primers used cence mounting medium. Fluorescence images were acquired on an for PCR included: epifluorescence microscope (BX61 attached to a DP71 camera, 5aF1 (forward)—ACAGCCTCTACCATCTCTCCACC Olympus) or a confocal laser scanning microscope (LSM 780, Zeiss). 3aR4 (reverse)—AACACGAGTGAGTCACCTTCGC Images were assembled and minimally adjusted in brightness and

The wild-type NaV1.7 allele and TAP-tagged NaV1.7 allele gave a contrast using Adobe Photoshop Elements 10. Bregma levels and 170-bp band and a 411-bp band, respectively. The TAP-tagged brain regions were identified according to the stereotaxic coordi-

NaV1.7 mRNA was confirmed by RT–PCR. Briefly, total RNA was nates in “The Mouse Brain” atlas by Paxinos and Franklin (2001). extracted from dorsal root ganglia with Qiagen RNEasy kit (Qiagen) The immunohistochemistry experiments in spinal cord were and 1.0 lg was used to synthesise cDNA using Bio-Rad iScript performed as described previously (Zhao et al, 2010) using the follow- cDNA synthesis kit with oligo-dT primers. The following primers ing primary antibodies: anti-FLAG antibody (1:400; F1804, Sigma),

were used to detect NaV1.7 mRNA: anti-CGRP antibody (1:100; #24112, ImmunoStar), anti-substance P

440 The EMBO Journal Vol 37 |No3 | 2018 ª 2018 The Authors Published online: January 15, 2018

Alexandros H Kanellopoulos et al NaV1.7-interacting proteins The EMBO Journal

(1:100; #20064, ImmunoStar) and biotin-conjugated isolectin B4 Western blot (1:200; L2140, Sigma). Proteins for Western blot were isolated from freshly excised dif- Behavioural analysis ferent tissues, such as DRG, spinal cord, sciatic nerve, olfactory bulb, cortex, hypothalamus, cerebellum, skin, lung, heart and

All behavioural tests were approved by the United Kingdom Home pancreas taken from TAP-tagged NaV1.7 mice and littermate control Office Animals (Scientific Procedures) Act 1986. Age (6–12 weeks)- mice. The protein samples were prepared the same as described in matched KI mice (four males and three females) and littermate wild- the section of single-step and tandem affinity purification in Materi- type (WT) controls (three males and four females) were used for als and Methods. Briefly, proteins extracted from different tissues acute pain behaviour studies. The experimenters were blind to the were homogenised in 1% CHAPS lysis buffer. genetic status of test animals. The Rotarod, Hargreaves’, von Frey The nuclear fraction and cell debris were removed by centrifuga- and Randall–Selitto tests were performed as described (Zhao et al, tion at 20,000 g for 15 min at 4°C. Protein concentrations were 2006). We used the same set of mice to perform acute behavioural determined with Pierce BCA protein assay kit, and then samples of tests. The order was Rotarod, von Frey, Hargreaves and Randall– 40 lg were separated on SDS–PAGE gel in Bio-Rad Mini-PROTEAN Selitto tests. We left one-day gap between Rotarod, von Frey and Vertical Electrophoresis Cell System and blotted to the Immobilin-P Hargreaves test, and 3-day gap between Hargreaves test and membrane (IPVH00010, Millipore) in transfer buffer (25 mM Randall–Selitto test. The formalin test was carried by intraplantar Tris–HCl, pH 8.3, 192 mM glycine, 0.1% SDS and 20% methanol) injections of 20 ll of 5% formalin. Age (8–12 weeks)-matched for 1 h at 100 V with a Bio-Rad transfer cell system. The membrane seven male KI mice and 10 male littermate wild-type (WT) controls was blocked in blocking buffer [5% nonfat milk in PBS–Tween were used. The mice were observed for 45 min and the time spent buffer (0.1% Tween 20 in 1× PBS)] for 1 h at room temperature biting and licking the injected paw was monitored and counted. and then incubated with primary antibody anti-FLAG (1:1,000; Two phases were categorised, the first phase lasting 0–10 min and Sigma, catalog #F1804) and anti-HAT (1:400; LSBio, #LS-C51508) in the second phase 10–45 min. All data are presented as blocking buffer overnight at 4°C. The membrane was washed three mean SEM. times with TBS–Tween (20 mM Tris, 150 mM NaCl, 0.1% Tween 20, pH 7.5) and then incubated with secondary antibody goat Single-step and tandem affinity purification anti-mouse or goat anti-rabbit IgG-HRP (1:4,000; Jackson Immuno- Research Laboratories) in TBS–Tween at room temperature for 2 h. In each round of sample preparation for single-step and tandem Detection was performed using a Western Lightning Chemilumines- affinity purification, DRG, olfactory bulbs, spinal cord and cence Reagent (Super Signal Western Dura, Thermo Scientific, hypothalamus samples were homogenised using Precellys Tissue #34075) and exposed to BioMax film (Kodak). Other primary Homogenizer and Precellys lysis kit (Precellys ceramic kit antibodies used for Western blotting were ankyrin-3 (Santa Cruz, 1.4 mm, Order no. 91-PCS-CK14, Peqlab) in 1% CHAPS lysis SC-12719), neurotrimin (Santa Cruz, SC-390941), Kif5b (Santa Cruz, buffer (30 mM Tris–HCl pH 7.5, 150 mM NaCl, 1% CHAPS, 1 SC-28538), PEBP1 (Thermo fisher, 36-0700), neurofascin (Abcam, complete EDTA-free protease inhibitor cocktail (Roche) in 10 ml ab31457), neuroligin (Abcam, ab177107) and AKAP12 (Abcam, of CHAPS lysis buffer) and further homogenised using an insulin ab49849). Concentrations were applied as suggested by the syringe. The lysates were incubated shaking horizontally on ice manufacturers. and clarified by centrifugation at 14,000 g for 8 min at 4°C. Protein concentrations were measured using the Pierce BCA Plasmids Protein Assay Kit (Product no. 23225, Thermofisher), and a total starting amount of 10 mg of protein containing supernatant was The following plasmids were obtained from Addgene (Cambridge, incubated with anti-FLAG M2 Magnetic beads (M8823, Sigma). MA): Tmed10 (TMED-BIO-HIS, #51852), Lat1 (pEMS1229, The coupling was carried out for 2 h at 4°C using an end-over- #29115). Gprin1 plasmid was obtained from OriGene

end shaker. Magnetic beads were collected on a DynaMAG rack (#RC207340). In order to perform in vitro validation of NaV1.7 (Invitrogen) and washed three times in 1% CHAPS Buffer and 1× interaction with synaptotagmin-2, we first cloned the human gene AcTEV protease cleavage buffer (50 mM Tris–HCl pH 8.0, for insertion into a mammalian expression plasmid. Synaptotag-

0.5 mM EDTA, 1 mM DTT). Bead-captured NaV1.7 TAP-tag min-2 was cloned from human dorsal root ganglion neuronal complex was released from the beads by incubation with AcTEV tissue. Whole mRNA was first reverse-transcribed into a cDNA protease (#12575015, Invitrogen) at 30°C for 3 h, finalising the library for PCR. Following insertion into a TOPO vector, this was single-step purification. For the tandem affinity purification, then used as the template to create the synaptotagmin-2 insert, protein eluates were collected after AcTEV cleavage and 15× which was then successfully cloned into a pcDNA3.1 IRES-AcGFP diluted in protein binding buffer (50 mM sodium phosphate, plasmid using the Gibson assembly method. The collapsin 300 mM NaCl, 10 mM imidazole, 0.01% Tween 20; pH 8.0). Response Mediator Protein 2 (Crmp2) gene was cloned from

AcTEV-cleaved NaV1.7 and its complex were then captured using human dorsal root ganglia neuron mRNA and cloned into a Ni-NTA beads (#36111, Qiagen). Ni-NTA beads were washed pcDNA3.1 plasmid using Gibson assembly. The Scn3b mamma- three times in protein binding buffer and incubated on an end- lian-expressing vector was used to investigate loss of function of

over-end shaker overnight at 4°C. TAP-tagged NaV1.7 protein NaV1.7 (Cox et al, 2006). The HA-tagged CaV2.2 was a gift from complexes were released from the Ni-NTA beads by boiling in 1× Prof. Annette Dolphin, and was co-expressed with the auxiliary SDS protein sample buffer. subunits, alpha2delta-1 and beta1b.

ª 2018 The Authors The EMBO Journal Vol 37 |No3 | 2018 441 Published online: January 15, 2018

The EMBO Journal NaV1.7-interacting proteins Alexandros H Kanellopoulos et al

Co-immunoprecipitation particle size, Waters) with a 180-min gradient of 3–40% solvent B

(solvent A: 99.9% H2O, 0.1% formic acid; solvent B: 99.9% ACN, Mouse tissues used in Co-IP experiments were chosen on the basis 0.1% formic acid). The Waters nanoAcquity UPLC system (final

of its known NaV1.7 expression in olfactory bulb, hypothalamus, flow rate, 250 nl/min) was coupled to a LTQ Orbitrap Velos spinal cord (lumbar enlargement), dorsal root ganglia (all cervical, (Thermo Scientific, USA) run in positive ion mode. The MS survey thoracic, lumbar and sacral DRGs) and sciatic nerve (both sides). scan was performed in the FT cell recoding a window between 300 All tissues from individual mice were either pooled or treated as and 2,000 m/z. The resolution was set to 30,000. Maximum of 20 individual tissue samples depending on experimental requirements. MS/MS scans were triggered per MS scan. The lock mass option Tissues were flash-frozen in dry ice immediately following dissec- was enabled, and polysiloxane (m/z 371.10124) was used for inter- tion and stored at 80°C to avoid protein degradation. HEK293 cells nal recalibration of the mass spectra. CID was done with a target

stably expressing TAP-tagged NaV1.7 were harvested by trypsinisa- value of 30,000 in the linear ion trap. The samples were measured tion followed by centrifugation at 800 rpm for 5 min. Cell pellets with the MS setting charge state rejection enabled, and only more were stored at 80°C to avoid protein degradation. Samples were than 1 charges procures ions selected for fragmentation. All raw MS lysed in a 1% CHAPS lysis buffer containing a protease inhibitor data were processed to generate MGF files (200 most intense peaks) cocktail and homogenised using ceramic zirconium oxide mix beads using the Proteowizard v.2.1.2476 software. The identification of of 1.4 and 2.8 mm lysing kit and homogeniser (Precellys). A total of proteins was performed using MGF files with the central proteomics 10–25 mg of protein was incubated with anti-FLAG M2 magnetic facilities pipeline. Mus musculus (Mouse) database containing beads for 2 h at 4°C. The bead–protein complex was then washed entries from UniProtKB was used in CPF Proteomics pipeline for with three cycles of five resin volumes of 1% CHAPS buffer and data analysis. This pipeline combines database search results from once with TEV–protease buffer (Invitrogen). The tagged protein was three search engines (Mascot, OMSSA and X!tandem k-score). The cleaved from the beads by the addition of TEV protease enzyme search was carried out using the following parameters. Trypsin was (Invitrogen) and incubated for 3 h at 37°C to elute the protein the enzyme used for the digestion of the proteins, and only one complex. Sample eluate was then separated from the beads and missed cleavage was allowed. The accepted tolerance for the precur- stored at 80°C until used for either Western blotting or mass spec- sor was 20 ppm and 0.5 Da for the fragment. The search encom- trometry. Co-immunoprecipitation in HEK293 cells was performed passed 1+,2+ and 3+ charge state, fixed modification for cysteine after transfection of the construct with the gene of interest into the carbamidomethyl and variable modification for asparagine and

TAP-tagged NaV1.7 stable cell line using a standard lipofectamine glutamine deamidation, and methionine oxidation. All trypsin frag- protocol (Lipofectamine 200, Invitrogen, #52887). All cells were left ments were added to an exclusion list. False discovery rate was 48 h after transfection before use in experiments. calculated by peptide/proteinprophet or estimated empirically from decoy hits a 1% FDR cut-off was used to filter identified proteins. Liquid chromatography–tandem mass spectrometry analysis The label-free analysis was carried out using the normalised spectral index (SINQ; Trudgian et al, 2011). The mass spectrometry proteo- Proteins cleaved from anti-FLAG M2 magnetic beads after affinity mics data have been deposited to the ProteomeXchange Consortium purification were tryptic-digested following a FASP protocol (http://www.proteomexchange.org/) via the PRIDE (Vizcaı´no et al, (Wisniewski et al, 2009). In brief, proteins were loaded to 30-kDa 2016) partner repository with the dataset identifier PXD004926. filters (Millipore) and then filter units were centrifuged at 14,000 g for 15 min to remove other detergents. Two hundred ll of urea Electrophysiology and patch clamp recordings buffer (10 mM dithiothreitol 8 M urea (Sigma) in 0.1 M Tris–HCl, pH 8.5) was added to the filters and left at room temperature for 1 h Whole-cell patch clamp recordings were conducted at room to reduce proteins. Then, filters were centrifuged to remove dithio- temperature (21°C) using an AxoPatch 200B amplifier and a Digi- threitol. Two hundred ll of 50 mM iodoacetamide (IAA) in urea data 1322A digitiser (Axon Instruments), controlled by Clampex buffer was added to filters and left for 30 min in the dark. Filters software (version 10, Molecular Devices). Filamented borosilicate were centrifuged as before to remove IAA. Then, the samples were microelectrodes (GC150TF-10, Harvard Apparatus) were pulled on buffer exchanged twice using 200 ll of urea buffer, and one more a Model P-97 Flaming/Brown micropipette puller (Sutter Instru-

time using 200 llof50mMNH4HCO3 in water. Forty ll of 50 ng/ll ments) and fire-polished to a resistance of between 2.5 and

trypsin in 50 mM NH4HCO3 was added to filter, filters were 4 MOhm. Standard pipette intracellular solution contained: 10 mM

vortexed briefly, and proteins were digested at 37°C for overnight. NaCl, 140 mM CsF, 1.1 mM EGTA, 1 mM MgCl2 and 10 mM After tryptic digestion, the filters were transferred to new collection HEPES. The standard bathing extracellular solution contained:

tubes and the peptides collected by placing the filter upside down 140 mM NaCl, 1 mM MgCl2, 3 mM KCl, 1 mM CaCl2 and 10 mM

and spinning. The samples were acidified with CF3COOH and HEPES. Both intracellular and extracellular solutions were adjusted desalted with C18 cartridge (Waters). The pure peptides were dried to a physiological pH of 7.3. The amplifier’s offset potential was by Speedvac (Millipore) and resuspended with 20 ll of 2% ACN, zeroed when the electrode was placed in the solution. After a giga- 0.1% FA. Five ll of samples was injected into Orbitrap Velos mass seal was obtained, short suction was used to establish whole-cell spectrometry (Thermo) coupled to a UPLC (Waters; The´ze´nas et al, recording configuration. Errors in series resistance were compen- 2013). sated by 70–75%. Cells were briefly washed with extracellular Analysis was carried out by nano-ultra-performance liquid chro- solution before a final 2 ml of solution was transferred to the dish. matography–tandem MS (nano-UPLC-MS/MS) using a 75 lm inner Cells were held at 100 mV for 2 min before experimental proto- diameter × 25 cm C18 nanoAcquity UPLCTM column (1.7 lm cols were initiated. Currents were elicited by 50-ms depolarisation

442 The EMBO Journal Vol 37 |No3 | 2018 ª 2018 The Authors Published online: January 15, 2018

Alexandros H Kanellopoulos et al NaV1.7-interacting proteins The EMBO Journal

steps from 80 mV to +80 mV in 5 mV increments. Compounds as described in the results or figure legends. The GraphPad Prism were added and mixed at the desired concentrations in extracellu- 6.0 was used to perform the statistical analysis. All data are lar solution before being added to the bath. Following addition of presented as mean SEM, and significance was determined at the compound, protocols were repeated on previously unrecorded P < 0.05. cells. All currents were leak-subtracted using a p/4 protocol. The following compounds were used in electrophysiology experiments: Expanded View for this article is available online. lacosamide ((R)-2-acetamido-N-benzyl-3-methoxypropionamide) was obtained from Toronto Research Chemicals Inc (L098500) and Acknowledgments tetrodotoxin was obtained from Sigma-Aldrich (T8024). Incubation We are grateful to Professor Seth GN Grant and Dr. Esperanza Fernandez with lacosamide was done for 5 h prior to recording. for the TAP-tag plasmids. We thank the Mass Spectrometry Laboratory of Voltage-clamp experiments were analysed using pCLAMP soft- Target Discovery Institute at University of Oxford for performing the mass ware and Origin (OriginLab Corp., Northampton, MA) software spectrometry analysis. We thank Dr. Massimo Signore who performed elec- programs. Current density–voltage analysis was carried out by troporation and blastocysts microinjection of ES cells in JP Martinez- measuring peak currents at different applied voltage steps and Barbera’s research group, ICH, UCL. This study was supported by the Well- normalised to cell capacitance (pA/pF). Voltage-dependent activa- come Collaborative Award 200183/Z/15/Z (to J.W., J.Z., J.C. and S.G.) and tion data were fitted to a Boltzmann equation y = (A2 + (A1 A2)/ Investigator Award 101054/Z/13/Z (to J.W.), the Medical Research Council (1 + exp((Vh x)/k)))*(x Vrev), where A1 is the maximal (MRC) Grant G091905 (to J.W.) and MRC CDA G1100340 (to J.C.), the amplitude, Vh is the potential of half-maximal activation, x is the Deutsche Forschungsgemeinschaft (DFG) Grants SFB 894/A17 (to F.Z.) and clamped membrane potential, Vrev is the reversal potential and k is PY90/1-1 (to M.P.). a constant. All Boltzmann equations were fitted using ORIGIN software. Author contributions

JZ generated the TAP-tagged NaV1.7 knock-in mouse. QM, SJG and JZ main-

NaV1.7 interaction protein selection and function analysis tained the mouse line. JZ and QM contributed to pain behavioural study and

analysed the data. JZ, MJ and JK generated the TAP-tagged NaV1.7 HEK293 cell Candidate proteins that may interact with Nav1.7 were selected line. AHK and JK performed the co-IP. HH performed mass spectrometry and by two criteria: (i) present in at least two knock-in biological analysed the data. MP, JZ and TM performed immunohistochemistry and anal- experiments but absent from wild-type experiments; (ii) present ysed the data. AHK, SL and JEL contributed to electrophysiology and analysed in more than three knock-in and more than one wild-type experi- the data. FZ, JJC, ACD, BMK and GB provided supervision. JNW and JZ conceived ments, the ratio of average abundance is more than 1.5-fold the study, designed the research and provided supervision. All authors increased in knock-in experiments as compared with wild-type prepared the manuscript. JZ and JNW wrote the article with inputs from AHK, experiments. Further cellular component and function classifi- HH and MP. cation were performed on PANTHER Classification System (11.0). Ingenuity pathway analysis (IPA) (QIAGEN) was used to elucidate Conflict of interest pathways and protein interaction networks using candidate The authors declare that they have no conflict of interest. proteins. References Southern blot analysis Angrand P-O, Segura I, Voelkel P, Ghidelli S, Terry R, Brajenovic M, Vintersten K, Klein R, Superti-Furga G, Drewes G, Kuster B, Bouwmeester T, Acker- The genomic DNA was extracted from either ES cells or tails of mice Palmer A (2006) Transgenic mouse proteomics identifies new 14-3-3- following the procedures as described (Sambrook & Russell, 2001). associated proteins involved in cytoskeletal rearrangements and cell The probes for Southern blot were amplified by PCR using mouse signaling. Mol Cell Proteomics 5: 2211 – 2227 genomic DNA isolated from C57BL/6 as a template and purified Bao L (2015) Trafficking regulates the subcellular distribution of voltage- with a Qiagen Gel Purification Kit. The restriction enzymes StuI, gated sodium channels in primary sensory neurons. Mol Pain 11: 61 BspHI and PsiI were used to digest genomic DNA for either wild- Bence M, Arbuckle MI, Dickson KS, Grant SG (2005) Analyses of murine type and knock-in bands. The sizes of wild-type and knock-in bands postsynaptic density-95 identify novel isoforms and potential translational are shown in Fig 2B. The primers used to create probes (50 external control elements. Mol Brain Res 133: 143 – 152 probe: 768 bp; 30 external probe: 629 bp) included: Black JA, Frézel N, Dib-Hajj SD, Waxman SG (2012) Expression of Nav1. 7 in 50PF (50 probe, forward)—ACCAAGCTTTTGATATCACCATCAT DRG neurons extends from peripheral terminals in the skin to 50PR (50 probe, reverse)—CAACTCGAGAACAGGTAAGACATGA central preterminal branches and terminals in the dorsal horn. Mol Pain 8: CAGTG 82 30PF (30 probe, forward)—TTTAAGCTTCCTGCCCCTATTCCTGCT Bonnon C, Wendeler MW, Paccaud J-P, Hauri H-P (2010) Selective export of 30PR (50 probe, reverse)—TTAGGATCCATGCACTACTGACTTGCT human GPI-anchored proteins from the endoplasmic reticulum. J Cell Sci TATAGGT 123: 1705 – 1715 Branco T, Tozer A, Magnus CJ, Sugino K, Tanaka S, Lee AK, Wood JN, Sternson Statistical analysis SM (2016) Near-perfect synaptic integration by Na v 1.7 in hypothalamic neurons regulates body weight. Cell 165: 1749 – 1761 Statistical analysis was performed using either repeated-measures Brittain JM, Duarte DB, Wilson SM, Zhu W, Ballard C, Johnson PL, Liu N, ANOVA with Bonferroni post hoc testing or unpaired Student’s t-test Xiong W, Ripsch MS, Wang Y (2011) Suppression of inflammatory and

ª 2018 The Authors The EMBO Journal Vol 37 |No3 | 2018 443 Published online: January 15, 2018

The EMBO Journal NaV1.7-interacting proteins Alexandros H Kanellopoulos et al

neuropathic pain by uncoupling CRMP-2 from the presynaptic Ca2+ differentially modulates their functional properties. J Neurosci 23: channel complex. Nat Med 17: 822 – 829 8261 – 8270 Cassidy JS, Ferron L, Kadurin I, Pratt WS, Dolphin AC (2014) Functional Ho C, Zhao J, Malinowski S, Chahine M, O’Leary ME (2012) Differential exofacially tagged N-type calcium channels elucidate the interaction with expression of sodium channel beta subunits in dorsal root ganglion auxiliary a2d-1 subunits. Proc Natl Acad Sci 111: 8979 – 8984 sensory neurons. J Biol Chem 287: 15044 – 15053 Catterall WA (2010) Signaling complexes of voltage-gated sodium and Isensee J, Krahé L, Moeller K, Pereira V, Sexton JE, Sun X, Emery E, Wood JN, calcium channels. Neurosci Lett 486: 107 – 116 Hucho T (2017) Synergistic regulation of serotonin and opioid signaling Chapman ER (2002) Synaptotagmin: a Ca2+ sensor that triggers exocytosis? contributes to pain insensitivity in Nav1. 7 knockout mice. Sci Signal 10: Nat Rev Mol Cell Biol 3: 498 – 508 eaah4874 Chen GI, Gingras AC (2007) Affinity-purification mass spectrometry (AP-MS) of Kanellopoulos AH, Zhao J, Emery EC, Wood JN (2017) Intracellular sodium serine/threonine phosphatases. Methods 42: 298 – 305 regulates opioid signalling in peripheral sensory neurons. bioRxiv https:// Chen-Izu Y, Shaw RM, Pitt GS, Yarov-Yarovoy V, Sack JT, Abriel H, Aldrich doi.org/10.1101/117952 [PREPRINT] RW, Belardinelli L, Cannell MB, Catterall WA, Chazin WJ, Keilhauer EC, Hein MY, Mann M (2015) Accurate protein complex retrieval by Chiamvimonvat N, Deschenes I, Grandi E, Hund TJ, Izu LT, Maier LS, affinity enrichment mass spectrometry (AE-MS) rather than Maltsev VA, Marionneau C, Mohler PJ et al (2015)Na+ channel function, affinity purification mass spectrometry (AP-MS). Mol Cell Proteomics 14: regulation, structure, trafficking and sequestration. JPhysiol593: 120 – 135 1347 – 1360 Klugbauer N, Lacinova L, Flockerzi V, Hofmann F (1995) Structure and Chi XX, Schmutzler BS, Brittain JM, Wang Y, Hingtgen CM, Nicol GD, Khanna functional expression of a new member of the tetrodotoxin-sensitive R(2009) Regulation of N-type voltage-gated calcium channels (Cav2. 2) voltage-activated sodium channel family from human neuroendocrine and transmitter release by collapsin response mediator protein-2 (CRMP- cells. EMBO J 14: 1084 2) in sensory neurons. J Cell Sci 122: 4351 – 4362 Koenig J, Werdehausen R, Linley JE, Habib AM, Vernon J, Lolignier S, Cox JJ, Reimann F, Nicholas AK, Thornton G, Roberts E, Springell K, Karbani G, Eijkelkamp N, Zhao J, Okorokov AL, Woods CG, Wood JN, Cox JJ (2015) Jafri H, Mannan J, Raashid Y, Al-Gazali L, Hamamy H, Valente EM, Gorman Regulation of Nav1.7: a conserved SCN9A natural antisense transcript S, Williams R, McHale DP, Wood JN, Gribble FM, Woods CG (2006)An expressed in dorsal root ganglia. PLoS One 10:e0128830 SCN9A channelopathy causes congenital inability to experience pain. Laedermann CJ, Cachemaille M, Kirschmann G, Pertin M, Gosselin RD, Chang Nature 444: 894 – 898 I, Albesa M, Towne C, Schneider BL, Kellenberger S, Abriel H, Decosterd I Cusdin FS, Clare JJ, Jackson AP (2008) Trafficking and cellular distribution of (2013) Dysregulation of voltage-gated sodium channels by ubiquitin ligase voltage-gated sodium channels. Traffic 9: 17 – 26 NEDD4-2 in neuropathic pain. J Clin Invest 123: 3002 – 3013 Dib-Hajj SD, Yang Y, Black JA, Waxman SG (2013) The Na(V)1.7 sodium Laedermann CJ, Abriel H, Decosterd I (2015) Post-translational modifications channel: from molecule to man. Nat Rev Neurosci 14: 49 – 62 of voltage-gated sodium channels in chronic pain syndromes. Front Dustrude ET, Wilson SM, Ju WN, Xiao YC, Khanna R (2013) CRMP2 protein Pharmacol 6: 263 SUMOylation modulates NaV1.7 channel trafficking. J Biol Chem 288: Leterrier C, Brachet A, Fache MP, Dargent B (2010) Voltage-gated sodium 24316 – 24331 channel organization in neurons: protein interactions and trafficking Emery EC, Luiz AP, Wood JN (2016) Nav1. 7 and other voltage-gated sodium pathways. Neurosci Lett 486: 92 – 100 channels as drug targets for pain relief. Expert Opin Ther Targets 20: Liu P, Jenkins NA, Copeland NG (2003) A highly efficient recombineering- 975 – 983 based method for generating conditional knockout mutations. Genome Res Fenalti G, Giguere PM, Katritch V, Huang X-P, Thompson AA, Cherezov V, Roth 13: 476 – 484 BL, Stevens RC (2014) Molecular control of d-opioid receptor signaling. Luo W, Wang Y, Reiser G (2011) Proteinase-activated receptors, nucleotide Nature 506: 191 P2Y receptors, and l-opioid receptor-1B are under the control of the type Fernandez E, Collins MO, Uren RT, Kopanitsa MV, Komiyama NH, Croning I transmembrane proteins p23 and p24A in post-Golgi trafficking. J MD, Zografos L, Armstrong JD, Choudhary JS, Grant SG (2009) Neurochem 117: 71 – 81 Targeted tandem affinity purification of PSD-95 recovers core McEwen DP, Isom LL (2004) Heterophilic interactions of sodium channel b1 postsynaptic complexes and schizophrenia susceptibility proteins. Mol Syst subunits with axonal and glial cell adhesion molecules. J Biol Chem 279: Biol 5: 269 52744 – 52752 Frank HY, Catterall WA (2003) Overview of the voltage-gated sodium channel Meadows LS, Isom LL (2005) Sodium channels as macromolecular complexes: family. Genome Biol 4: 1 implications for inherited arrhythmia syndromes. Cardiovasc Res 67: Ge X, Qiu Y, Loh HH, Law P-Y (2009) GRIN1 regulates l-opioid receptor 448 – 458 activities by tethering the receptor and G protein in the lipid raft. J Biol Minett MS, Nassar MA, Clark AK, Passmore G, Dickenson AH, Wang F, Chem 284: 36521 – 36534 Malcangio M, Wood JN (2012) Distinct Nav1.7-dependent pain sensations Habib AM, Wood JN, Cox JJ (2015) Sodium channels and pain. In Pain Control. require different sets of sensory and sympathetic neurons. Nat Commun 3: Handbook of Experimental Pharmacology, Schaible HG (ed.), pp 39 – 56. 791 Berlin: Springer Minett MS, Pereira V, Sikandar S, Matsuyama A, Lolignier S, Kanellopoulos AH, Hendrich J, Van Minh AT, Heblich F, Nieto-Rostro M, Watschinger K, Mancini F, Iannetti GD, Bogdanov YD, Santana-Varela S, Millet Q, Baskozos Striessnig J, Wratten J, Davies A, Dolphin AC (2008) Pharmacological G, MacAllister R, Cox JJ, Zhao J, Wood JN (2015) Endogenous opioids disruption of calcium channel trafficking by the a2d ligand gabapentin. contribute to insensitivity to pain in humans and mice lacking sodium Proc Natl Acad Sci 105: 3628 – 3633 channel Nav1.7. Nat Commun 6: 8967 Herzog RI, Liu C, Waxman SG, Cummins TR (2003) Calmodulin binds to Nahin RL (2015) Estimates of pain prevalence and severity in adults: United the C terminus of sodium channels Nav1. 4 and Nav1. 6 and States, 2012. J Pain 16: 769 – 780

444 The EMBO Journal Vol 37 |No3 | 2018 ª 2018 The Authors Published online: January 15, 2018

Alexandros H Kanellopoulos et al NaV1.7-interacting proteins The EMBO Journal

Namadurai S, Balasuriya D, Rajappa R, Wiemhofer M, Stott K, Klingauf J, Trudgian DC, Ridlova G, Fischer R, Mackeen MM, Ternette N, Acuto O, Kessler Edwardson JM, Chirgadze DY, Jackson AP (2014) Crystal structure and BM, Thomas B (2011) Comparative evaluation of label-free SINQ molecular imaging of the Nav channel beta3 subunit indicates a trimeric normalized spectral index quantitation in the central proteomics facilities assembly. J Biol Chem 289: 10797 – 10811 pipeline. Proteomics 11: 2790 – 2797 Namadurai S, Yereddi NR, Cusdin FS, Huang CL, Chirgadze DY, Jackson AP Vetrivel KS, Gong P, Bowen JW, Cheng H, Chen Y, Carter M, Nguyen PD, (2015) A new look at sodium channel beta subunits. Open Biol 5: 140192 Placanica L, Wieland FT, Li Y-M (2007) Dual roles of the transmembrane Nassar MA, Stirling LC, Forlani G, Baker MD, Matthews EA, Dickenson AH, protein p23/TMP21 in the modulation of amyloid precursor protein Wood JN (2004) Nociceptor-specific gene deletion reveals a major role for metabolism. Mol Neurodegener 2: 4 Na(v)1.7 (PN1) in acute and inflammatory pain. Proc Natl Acad Sci USA Vizcaíno JA, Csordas A, del-Toro N, Dianes JA, Griss J, Lavidas I, Mayer G, 101: 12706 – 12711 Perez-Riverol Y, Reisinger F, Ternent T (2016) 2016 update of the PRIDE Oeffinger M (2012) Two steps forward—one step back: advances in affinity database and its related tools. Nucleic Acids Res 44:D447 – D456 purification mass spectrometry of macromolecular complexes. Proteomics Wang Y, Brittain JM, Jarecki BW, Park KD, Wilson SM, Wang B, Hale R, Meroueh 12: 1591 – 1608 SO, Cummins TR, Khanna R (2010) In silico docking and electrophysiological Oka Y, Ye M, Zuker CS (2015) Thirst driving and suppressing signals encoded characterization of lacosamide binding sites on collapsin response mediator by distinct neural populations in the brain. Nature 520: 349 – 352 protein-2 identifies a pocket important in modulating sodium channel slow Ott S, Costa T, Herz A (1988) Sodium modulates opioid receptors through a inactivation. J Biol Chem 285: 25296 – 25307 membrane component different from G-proteins. Demonstration by target Weiss J, Pyrski M, Jacobi E, Bufe B, Willnecker V, Schick B, Zizzari P, Gossage size analysis. J Biol Chem 263: 10524 – 10533 SJ, Greer CA, Leinders-Zufall T, Woods CG, Wood JN, Zufall F (2011) Loss- Paxinos G, Franklin KB (2001) The mouse brain in stereotaxic coordinates. San of-function mutations in sodium channel Nav1.7 cause anosmia. Nature Diego, CA: Academic Press 472: 186 – 190 Ratcliffe CF, Westenbroek RE, Curtis R, Catterall WA (2001) Sodium channel Wildburger NC, Ali SR, Hsu W-CJ, Shavkunov AS, Nenov MN, Lichti CF, LeDuc beta 1 and beta 3 subunits associate with neurofascin through their RD, Mostovenko E, Panova-Elektronova NI, Emmett MR, Nilsson CL, Laezza extracellular immunoglobulin-like domain. J Cell Biol 154: 427 – 434 F(2015) Quantitative proteomics reveals protein–protein interactions with Rush AM, Cummins TR, Waxman SG (2007) Multiple sodium channels and fibroblast growth factor 12 as a component of the voltage-gated sodium their roles in electrogenesis within dorsal root ganglion neurons. J Physiol channel 1.2 (Nav1.2) macromolecular complex in mammalian brain. Mol 579: 1 – 14 Cell Proteomics 14: 1288 – 1300 Sambrook J, Russell DW (2001) Molecular cloning: a laboratory manual, 3rd Wilson SM, Khanna R (2015) Specific binding of lacosamide to collapsin edn. New York, NY: Cold Spring Harbor Laboratory Press response mediator protein 2 (CRMP2 ) and direct impairment of its Sampo B, Tricaud N, Leveque C, Seagar M, Couraud F, Dargent B (2000) canonical function: implications for the therapeutic potential of Direct interaction between synaptotagmin and the intracellular loop I-II lacosamide. Mol Neurobiol 51: 599 – 609 of neuronal voltage-sensitive sodium channels. Proc Natl Acad Sci 97: Wisniewski JR, Zougman A, Nagaraj N, Mann M (2009) Universal sample 3666 – 3671 preparation method for proteome analysis. Nat Methods 6: 359 Shao D, Baker MD, Abrahamsen B, Rugiero F, Malik-Hall M, Poon WY, Cheah Woodcock J (2009) A difficult balance — pain management, drug safety, and KS, Yao KM, Wood JN, Okuse K (2009) A multi PDZ-domain protein Pdzd2 the FDA. N Engl J Med 361: 2105 – 2107 contributes to functional expression of sensory neuron-specific sodium Yang L, Dong F, Yang Q, Yang PF, Wu R, Wu QF, Wu D, Li CL, Zhong YQ, Lu YJ, channel Na(V)1.8. Mol Cell Neurosci 42: 219 – 225 Cheng X, Xu FQ, Chen L, Bao L, Zhang X (2017) FGF13 selectively regulates Sheets PL, Gerner P, Wang C-F, Wang S-Y, Wang GK, Cummins TR (2006) heat nociception by interacting with Nav1.7. Neuron 93: 806 – 821 Inhibition of Nav1. 7 and Nav1. 4 sodium channels by trifluoperazine Zhao J, Seereeram A, Nassar MA, Levato A, Pezet S, Hathaway G, Morenilla- involves the local anesthetic receptor. J Neurophysiol 96: 1848 – 1859 Palao C, Stirling C, Fitzgerald M, McMahon SB, Rios M, Wood JN, London Srinivasan J, Schachner M, Catterall WA (1998) Interaction of voltage-gated Pain C (2006) Nociceptor-derived brain-derived neurotrophic factor sodium channels with the extracellular matrix molecules tenascin-C and regulates acute and inflammatory but not neuropathic pain. Mol Cell tenascin-R. Proc Natl Acad Sci 95: 15753 – 15757 Neurosci 31: 539 – 548 Terpe K (2003) Overview of tag protein fusions: from molecular and Zhao J, Lee M-C, Momin A, Cendan C-M, Shepherd ST, Baker MD, Asante C, biochemical fundamentals to commercial systems. Appl Microbiol Bee L, Bethry A, Perkins JR, Nassar MA, Abrahamsen B, Dickenson A, Cobb Biotechnol 60: 523 – 533 BS, Merkenschlager M, Wood JN (2010) Small RNAs control sodium Thézénas ML, Huang H, Njie M, Ramaprasad A, Nwakanma DC, Fischer R, channel expression, nociceptor excitability, and pain thresholds. J Neurosci Digleria K, Walther M, Conway DJ, Kessler BM (2013) PfHPRT: a new 30: 10860 – 10871 biomarker candidate of acute Plasmodium falciparum infection. J Proteome Res 12: 1211 – 1222 License: This is an open access article under the Toledo-Aral JJ, Moss BL, He Z-J, Koszowski AG, Whisenand T, Levinson SR, terms of the Creative Commons Attribution 4.0 Wolf JJ, Silos-Santiago I, Halegoua S, Mandel G (1997) Identification of License, which permits use, distribution and reproduc- PN1, a predominant voltage-dependent sodium channel expressed tion in any medium, provided the original work is principally in peripheral neurons. Proc Natl Acad Sci 94: 1527 – 1532 properly cited.

ª 2018 The Authors The EMBO Journal Vol 37 |No3 | 2018 445